www.delorie.com/archives/browse.cgi   search  
Mail Archives: pgcc/1999/08/04/02:13:30

Date: Mon, 2 Aug 1999 14:00:19 +0200 (MET DST)
From: Piotr Wasiewicz <pwasiewi AT elka DOT pw DOT edu DOT pl>
To: pgcc AT delorie DOT com
Subject: Segmentation fault in k6
In-Reply-To: <Pine.GSO.4.05.9907261231500.10311-100000@elektron.elka.pw.edu.pl>
Message-ID: <Pine.GSO.4.05.9908021352080.19395-100000@elektron.elka.pw.edu.pl>
MIME-Version: 1.0
Reply-To: pgcc AT delorie DOT com
X-Mailing-List: pgcc AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com

Hello,
	I have k6-2 processor. I compiled under Redhat 6.0 pgcc patched
egcs (with -06 and -mpentium). After compiling with it mbench (byte
benchmark for linux), I noticed that my programs run only compiled with
option -fforce_addr (I had to add "-mk6" also). 
	Program mbench originally compiled by authors was quicker in
integer parts, but have slower floating points parts than programs
compiled with my new egcs (-mk6 -O6 -fforce_addr).
	How to understand this strange fact that old gcc with libc5 has
better optimized integer parts. Maybe I need some special options?
	Write me about it.

With best regards,	
	Peter Wasiewicz  

X//N\\XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
///|\\\ Piotr Wasiewicz http://tichy.ise.pw.edu.pl/~pwas
///A\\\ Ph. D. Student from GAIR http://tichy.ise.pw.edu.pl/~gair
G-A*I-R   Inst. of Electronic Systems at the Warsaw Univ. of Tech.
\\\L///   00-665 Warsaw, Nowowiejska 15/19, Poland; pwas AT ise DOT pw DOT edu DOT pl
\\\|///   home phone: (+4822)6670827; fax: (+4822)8252300       
X\\M//XXXXMy UIN: 10145599 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

On Mon, 26 Jul 1999, Piotr Wasiewicz wrote:

> Hello,
> 	I used the latest Red Hat Linux 6.0 with updates and the pgcc
> compiler (first I uninstall egcs) from Foyer rpm packages. I compiled
> files with gcc and all is all right even with -O6 -mk6, but when I try to
> compile C++ files with g++ (even without any parameters) then obtained
> programs give "Segmentation fault" what is for me strange.
> 	After some debugging I noticed errors within file genops.c
> from distribution with IO_FILE data (some overflows and other errors).
> 	What should I do?
> 
> With best regards,
> 	Peter Wasiewicz
> 
> X//N\\XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
> ///|\\\ Piotr Wasiewicz http://tichy.ise.pw.edu.pl/~pwas
> ///A\\\ Ph. D. Student from GAIR http://tichy.ise.pw.edu.pl/~gair
> G-A*I-R   Inst. of Electronic Systems at the Warsaw Univ. of Tech.
> \\\L///   00-665 Warsaw, Nowowiejska 15/19, Poland; pwas AT ise DOT pw DOT edu DOT pl
> \\\|///   home phone: (+4822)6670827; fax: (+4822)8252300       
> X\\M//XXXXMy UIN: 10145599 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
> 

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019